Gajim - 2023-12-16

> Hi all ! does xmpp implements post-quantum cryptography yet ? no... maybe wait until we have a functioning quantum computer... ↺

instead of fearmongering

but its not "XMPP" you would need to ask which servers would support it and clients, but thats mostly down to the TLS implementation

> Is the the point when XMPP will gain broad adoption? ;) I wish... everyone tells me "gajim looks like ass" ↺

Gajim functions better than most chat platforms... yet people only care about eye candy

polarian: what's "not ass" for them?

> polarian: what's "not ass" for them? Discord ↺

Then you're talking with scarabs.

Apps should look nice by respecting the system theme, and the system theme should be nice to provide eye candies everywhere, not just in certain non-native apps.

polarian: > no... maybe wait until we have a functioning quantum computer... > > instead of fearmongering its best to implement resistant cryptography before we need it. once we have "functioning" quantum computers, itll be too late by then

amogus, that's nothing Gajim specific though, and Gajim would not be a place to start

true, it would be better to update OMEMO or add a new crypto for encryption, and then implement it

Some time ago, I tried out Dino. IIRC, Conversations could not read/decrypt OMEMO mesasges received from Dino due to Dino using an upgraded OMEMO spec. Not sure if Gajim can.

is there a breaking change coming in future OMEMO spec? any ideas?

kurion, afaik there is no issue between Gajim and Dino

Only clients implementing OMEMO 0.8+ will have a different mode of encryption.

It's probably not important but I think it would be nice if instead of "A setting not related to privacy has been changed" would instead tell us exactly what changed

I guess it's just a matter of the extra code that translates each change into a sentence for humans?

Rather than making sentences, we could simply print a message of '@var = <value/>'

For example 'Room settings changed: muc#roomconfig_allowinvites = 1'

fjklp, I would not expose protocol to users. But yes, displaying individual messages sounds better

True, exposing the protocol is probably not best for many users.

I do see it as a goal to make it less necessary to understand how xmpp works in order to use it

I was just thinking of a way to write less code to get the end result

Now this has me wondering if there is a canonical friendly description of these protocol features

cal0pteryx: concealing XMPP-specific concepts from users makes it harder to get how does it differ from other IMs and what its benefits are.

bodqhrohro: some people want things to "Just Work", and that's not necessarily something to fight against

there's an ever-present tension there

fjklp: that's an unrealistic assumption.

The more you require a user to learn, the more resistance there is to them switching to your thing

The more users expect things to work well, the more butthurt they have when they break. And they eventually break, inevitably.

So, deliver a broken product to reduce expectations, and therefore reduce disappointment? lol

Preemptive brokenness. It's a feature, not a bug.

fjklp: how honesty and verbosity are a brokenness lol?

look at discord. It just works.

fjklp: it doesn't.

people use it. people like it.

Furthermost, it's greedy.

irrelevant

we're talking about UI/UX

Their crazy spam detector bans innocent accounts.

However, I'm not saying we should exactly try to make a discord clone. Discord would never show messages like "Room settings have changed"

innocent: see, that's one of XMPP-specific things many newcomers don't realize.

> Their crazy spam detector bans innocent accounts. You tagged me or did my client automatically tagged me because I of my name? ✏ ↺

They attempt to @mention as in Twitter.

Despite XMPP is much older than Twitter lol.

And has its own conventions.

And a large userbase which knows it since 00s.

No need to make it something different for the sake of attracting the audience from other IMs.

Skype has went through this so only the brand has left, where is it now? Before Microsoft it's power was P2P sending of files of arbitrary size, now it's a typical cloud-based messenger with no significant benefits over other cloud-based messengers, they're perpetually behind others in terms of features.

XMPP should stand forward by things present in nothing else mainstream, otherwise there would be just no reason to switch if it would be just a poor clone of something else.

I already mentioned real-time texting as an example, yet it's already was implemented in early ICQ and forgotten since then, it might be cool and attractive for those who don't remember that lol.

@username instead of *username* would be better.

innocent: it already works how it does, no need to break compatibility.

> innocent wrote: > @username instead of *username* would be better. Nah.

@username is more intuitive, I think many will agree on this

> innocent: it already works how it does, no need to break compatibility. Yeah. ↺

Typing just the name is enough of a mention.

"Accidental" highlighting was a neat way of trolling back then actually.

kurion: nothing intuitive if you have not seen it elsewhere.

But it may cause confusions if the message has username.

To me @ would be a nice shortcut to initialise name selection of curren chat's participants. Like a small popup textarea where as I type I can enter very quickly someone's name.

That's intentional, you can set a nickname "the" and express a butthurt when someone mentions you all *the* time lol.

Many clients support custom words to be highlighted for besides of the nickname, natively or via plugins.

For example, for tracking topics of interest.

> if you have not seen it elsewhere. true. but doesn't @username come off as "this message is intended to be sent to the address of username"?

Geld: tabstopping works faster than some kind of popups, and are prone to focus loss.

Popup widgets which interrupt the workflow and may steal the focus or cause interruptions are really annoying for touch typists.

> kurion wrote: >> if you have not seen it elsewhere. > true. but doesn't @username come off as "this message is intended to be sent to the address of username"? That's Discord way of doing it.

The neat official Discord's feature missing in Gajim is Ctrl+K lol.

For quick switching between chats and whatever.

It's present in Nheko, BTW.

> cal0pteryx: concealing XMPP-specific concepts from users makes it harder to get how does it differ from other IMs and what its benefits are. I did not say that. I said I don't want to expose protocol, i.e. protocol strings. I also said displaying individual, i.e. specific messages would be good.

> true, it would be better to update OMEMO or add a new crypto for encryption, and then implement it What are you hiding which is so important that we need to get super paranoid about post-quantum?

governments and companies will not finish quantum before they have quantum resistance

they need to protect their intellectual property ✏

people need to take a chill pill for a moment and focus on making XMPP more functional, not start stressing over "what if quantum breaks the encryption"

> No need to make it something different for the sake of attracting the audience from other IMs. I'd like a syntax that allows me to say a user's name without pinging them, as well as the option to ping them. It's useful.

> Their crazy spam detector bans innocent accounts. me

I type fast

fuck them anyways, not like I want to use Discord but when so many open source projects are now on Discord (only) its almost impossible not to

> @username instead of *username* would be better. No... its so much nicer without the @, don't like it, go back to Discord 🙃

fjklp, nah.. IRC is the same

and IRC is holy ✏

polarian, please leave out the profanity

fjklp: > I'd like a syntax that allows me to say a user's name without pinging them Impossibru, there in any way are legacy clients which would highlight them on bare substring, changing that in some of clients would only bring confusion.

> governments and companies will not finish quantum before they have quantum resistance you don't seem to recognize that breaking encryption will allow decryption of stored old messages. You want to implement it as soon as possible, even if quantum breaking won't happen for decades. Still, not a gajim problem.

Different defaults for nickname suffix (: vs ,) were enough of confusion already.

(does anything but Gajim actually use `,`, huh?)

> fjklp, nah.. IRC is the same How is that at all relevant? I'm not suggesting that xmpp should be an extension of irc.

> polarian, please leave out the profanity Apologies

bodqhrohro, nothing breaks if a "mention" standard is implemented by newer clients, while old clients still mention on exact matches with the username (while newer simply don't)

> you don't seem to recognize that breaking encryption will allow decryption of stored old messages. You want to implement it as soon as possible, even if quantum breaking won't happen for decades. Still, not a gajim problem. I do realise that, but that means as soon as TLSv1.3 is broken

these messages are open

why worry about the inevitable

are you serious...

I'm done

cal0pteryx: the point is to explicitly avoid mentions, I told that's impossible because there's no guarantee that receiving buddy's clients wouldn't highlight by bare substring anyway.

it is possible for the clients that implement the mention standard

fjklp, so you are saying we should all put tin foil hats on and worry about how anything we send today could be compromised 10 years down the line?

then old clients continue to work as they always did

a lot of post crypto algorithms are not highly tested yet

they are experimental

fear mongering is not going to fix the problem

and you can say "make OMEMO post quantum"

but you say that like its as easy as that

Prioritizing forward security is not paranoia, nor is it fear mongering. This topic is dead.

fjklp, and you are someone who can't take the fact others have differing opinions...

using experimental algorithms could end up to data being compromised even faster

_could_

> cal0pteryx: the point is to explicitly avoid mentions, I told that's impossible because there's no guarantee that receiving buddy's clients wouldn't highlight by bare substring anyway. bodqhrohro, And how would you ensure that clients would properly highlight an mention with an @ in it?

You never know

they could be using any client

Using a non-ASCII nickname is another way of trolling touch typists for a long time.

bodqhrohro, and whats your suggestion, sitting an @ at the beginning and calling it a day?

polarian: some newcomers already do.

Nothing to be done really.

They are just Discord users, they will figure out how to do things eventually

I'm not sure if that's not a default already in Beagle.IM lol.

polarian: why only Discord? It's a widespread convention among modern messengers.

Twitter has pioneered it, as I told.

Others just blindcopied and eventually it became a new standard de-facto.

Since there is no standard yet, Gajim won't change anything

> yet Sounds terrifying.

bodqhrohro, ✏

thats where most edgy teens are from 🙃

bodqhrohro: ok bro, can you code that?

Geld: code what?

@

what is the point!!!

Functionality in Gajim

it depends on the client

I still can't even run freshest Gajim probably lol.

So since you can't contribute anything to the code and since your idea has been negatively received by everyone, please move on.

Last commit working for me is 24548c14e87e5044eebaf0092ab27ef19cfb8659, before the OMEMO support being merged into the main codebase.

Geld: what idea?

K I messed up

Whoever who wanted @username plz stfu thank you for attention lol 😆 ✏

Whoever wanted @username plz stfu thank you for attention lol 😆 ✏

Actually, tabstopping does not work with @ as of yet.

fjklp: the protocol does not tell us what changed, regarding the privacy message

OMEMO carbons are broken for me atm. In both directions, Gajim 1.8.4 on both devices. How can I accept my own keys, again?

> fjklp: the protocol does not tell us what changed, regarding the privacy message then how do you detect that there was a change to present? you just go by the fact that we receive a status code 104? Even if this, gajim can know what changed if it caches the current variables and checks what changes when it does a discoinfo. But it would also be nice if the server tells us what changes. ✏

Also, how do we know that the setting was "not related to privacy"?

it does not tell us what changed

it tells us that a setting regarding privacy was changed

and gajim tells exactly that to the user

would be a diff nice, yeah, but the server does not give us one, and i cannot think of a trivial implementation to diff 2 disco info and translate that into sentences which make sense to the user

I guess we go by status code 104 to know it's non-privacy-related

yes

However, afaict, we could track changes manually, maybe you don't care to ✏

much work little gain

you can simply ask the moderator what was changed

or look if anything important to you changed

Why do we need a rotary phone, just call an operator and ask! Aren't you overly introverted to talk to the operator, right? ✏

bodqhrohro, you're not helpful

cal0pteryx: I'm not supposed to be helpful, I'm not a phone operator.

Or were they called dispatchers?

Out of curiosity, say I submitted a MR for this functionality, would that be viewed as creating undesired compexity?

Where can I find the chat history of all contacts and muc and see/export/delete it? I do not find it anymore in the "new" GUI...

Roi, you find them in each individual chat (try searching from the search bar)

exporing is done via accounts > your account > privacy > export

Thank you. I'm a step closer. :-)

Is there a way to export just a single muc?

Or contact?

exporting: no. removing: yes (via chat details)

Okay thank you. Too bad.